Abstract
INTRODUCTION: The proximity of the rectum to the prostate in radiation therapy for prostate cancer presents a significant dosimetric challenge, leading to high rectal doses and resulting in detrimental side effects. Perirectal tissue spacing reduces rectal dose and gastrointestinal toxicities by mechanically separating these organs. We retrospectively compared balloon and PEG hydrogel spacers, focusing on spacer geometry, symmetry, rectal dosimetry, and GI/GU toxicity. METHODS: Sixty-seven men with localized prostate cancer treated with EBRT were analysed (balloon = 33; PEG hydrogel = 34). Symmetry was graded on axial CT at apex, mid-gland, and base with a five-tier midline scale (SYM-1 = optimal). Anteroposterior, laterolateral, and craniocaudal separations were measured. Rectal V60%-V100% were taken from dose-volume histograms. Acute (≤90 d) and late (>90 d) GI/GU toxicities were scored (CTCAE v4.0). Two-sided p ≤ 0.05 was significant. RESULTS: Optimal symmetry occurred in 33% (balloon) vs 14% (PEG hydrogel); asymmetry SYM-4/5 in 27% vs 24% (p = 0.21). At the apex, balloon spacers consistently created measurable separation, whereas 3 patients (9%) with PEG hydrogel demonstrated complete absence of spacing. Mean anteroposterior separation was larger with balloon at all levels (p < 0.001). Laterolateral differed inferiorly (2.4 cm vs 1.9 cm; p = 0.01). Craniocaudal length averaged 4.8 cm vs 4.3 cm (p < 0.001). Rectal V60-V100% showed no significant differences. Acute toxicity was low: GI grade 1 in 6% (balloon) vs 0%, with one grade 3 GI in PEG hydrogel; GU grade 1 in 13% vs 29%, grade 2 in 10% vs 7%. Late events: GI grade 2 in 0% vs 7%; GU grade 3 in one patient per cohort (~3%); other late toxicities mild and similar. CONCLUSION: The balloon spacer achieved greater, more uniform separation including improved apical symmetry, and showed fewer early GI events and lower mild acute GU rates, while rectal doses remained comparable. Prospective studies with longer follow-up are needed to confirm long-term benefit.